Real Time Information Director User Documentation

RTID Metadata Language
Hewlett-Packard Company 66 529618-002
public abstract class Library
The next few lines of metadata define values for partitioning parameters. In this case, the
data will be divided across four partitions, starting with partition 0. You need not use the
same number of partitions for all the tables in your solution, but the Library application
does so for simplicity.
The partition range actually specifies the number of keys available for assignment to the
partition. In this case, each partition can have a million keys assigned to it.
The block size is the number of keys that will be allocated at once. Only one key is
needed for each document, of course, but a PartitionedKeyDispenser class fetches a block
of keys at a time; any other instance of the Director would have to start with the next set
of that many keys.
{
public static final short STARTING_PARTITION = 0;
public static final short NO_OF_PARTITIONS = 4;
public static final long PARTITION_RANGE = 1000000;
public static final int KEYS_BLOCK_SIZE = 100;
public static final PartitionParameters partitionParameters =
new
PartitionParameters("LAB",STARTING_PARTITION,NO_OF_PARTITIONS,PARTITION_
RANGE);
The round-robin partitioning strategy is used for various tables within the example and
defined once here:
public static final PartitioningStrategy roundRobinStrategy =
new RoundRobinPartitionId(STARTING_PARTITION,
NO_OF_PARTITIONS);
The next few lines define an EnrichedRecord object based on the BRANCH table within
the library example:
public static EnrichedRecord branchRecord =
new EnrichedRecord("BRANCH");
The BRANCH table is used to map between the name of a branch library and the internal
identifier for the branch, assigned by the Director. The following statement describes
how the branchMapper correlates these elements; it would be required, even if the data
were not being partitioned:
public static final Mapper branchMapper =